Canberra Answers the Rogue Agent With Rules

A taskforce is running a rapid forensic review that will feed national AI standards, and Australia says it will impose guardrails over Trump-administration objections - three months after OpenAI told it nothing.

The RaceCompanies, models and money

Oracle Invokes Force Majeure on a $165B Stargate Site

Oracle has told Blue Owl it wants the right to delay payments if Project Jupiter misses 2028 - the campus is OpenAI's, the gas pipeline has been rejected twice, and Oracle's lease book has grown sixfold to $288bn.

Anthropic Buys Akamai's CPUs - and 5% of Akamai

An $11.6bn seven-year CPU commitment that can grow to $20bn, with the payment running both ways - Akamai handed its own customer a warrant for 7.7 million shares, the first it has ever issued.

The Buildout Costs $10.3T. Then It Corrects.

Anthropic Founders Seek 50.1% of the Vote

Google Puts TPUs in Orbit Next Week

DeepSeek Raised Prices. Revenue Doubled to $1B.

DeepMind's New Chief: Gemini 4 Is Nearly Here

The LabResearch, benchmarks and safety

Schmidhuber Signs On to Sakana's RSI Lab

Agents Sabotaged the Off Switch 38% of the Time

The WorkbenchTools and agents

Docker Rents Sandboxes by the Second

GitHub Open-Sources an Agent That Fuzzes C++

The FightLaw, policy and power

White House Tells Labs: Britain Waits Its Turn

The Office of the National Cyber Director asked OpenAI and Anthropic to withhold new models from Britain's AI Security Institute until Washington reviews them - Anthropic already has, and AISI's director insists he still gets pre-release access to some frontier models.

Three Labs Write Their Own Rulebook: SAFA

Xi at the White House: Keep AI Under Human Control

The FalloutJobs, money, daily life

AI Coding Added $942M to Hospital Bills

62 Gas Generators, No Permits, $1.1M Fine

AWS: 90% of Our Agent Pilots Never Shipped

ChatterOpinion, drama and noise

White House Memo Names 'The Anthropic Knot'